The hardest part of AI was never the intelligence — it was the context. A model can ace a benchmark and still fail the moment it meets a real enterprise: its messy data, its tribal knowledge, its accountability, its consequences. Pangea calls this the Context Paradox — AI that's brilliant in theory and useless in practice. Closing that gap, and making AI genuinely usable by real people, is the whole job.
We build clarity, not just code.
How it started
It started, as the best companies often do, with a shared frustration. On appraisal day — June 4, 2019 — over drinks in a dim-lit pub, a small group of engineers asked the question that doesn't go away once you've said it out loud: what if we started something of our own?
By that October, surrounded by sticky notes and whiteboards, the frustration had sharpened into a thesis. They kept hitting the same wall in their day jobs: why is adopting AI so complex, so opaque, so hard to trust? Enterprises were pouring money into pilots and watching most of them stall — not because the models were weak, but because nobody could see inside them, and nothing was built for the context the business actually lived in.
That was the spark. Pangea was founded in 2020 to make AI something a business could actually understand and rely on — transparent instead of black-box, explainable instead of magic, built for the people who'd have to live with its decisions. Three founders started the company out of a tiny room in Bangalore, fueled by big ideas and an unreasonable amount of chai. The bet was simple and stubborn: treat clarity, ethics, and trust as features, not afterthoughts — and build AI that holds up where it's hardest, inside real enterprise workflows. I joined full-time in 2023, as CTO, and later became a co-founder myself.
What Pangea built
The work fell into two distinct phases, split by the moment ChatGPT landed.
Phase one — explainable AI
Explainable AI was what we were aiming for, and we built a product around it: TuringXai. The feedback from the people we showed it to was genuinely interesting — a couple of resellers even wanted to try selling it. But overall, I think it missed the convincing factor: what it actually meant for a business. We kept improving it, and ultimately shipped a couple of customised versions for clients. That was when we knew it had run its course.
Coincidentally, right around then, a company called OpenAI released ChatGPT — and it was something new.
Phase two — generative AI, and everything around it
We had worked with AI/ML models well before ChatGPT, but this looked like a direction worth trying. Our first product in the generative AI space was an image generation and editing tool. The idea was solid, we had a few competitors, and we certainly made some mistakes. Ultimately the product didn't work — but we didn't give up.
By that point our strength wasn't only AI. We had built deep technology expertise, and we worked across some very diverse industries building domain-specific products for them.
Over the next three years we had multiple clients and multiple successful solutions — but we still hadn't found success with a homegrown product. That's when we decided we needed a reset: pick a domain, pick a use case, and specialise. That's how Antesian Software Labs was born.
We still don't have it all figured out for Antesian, and we're in no hurry to. But when we do, the focus will be razor sharp.
The journey
| Year | Milestone |
|---|---|
| Jun 2019 | The spark — a shared frustration, a pub, "what if?" |
| Oct 2019 | The realization — sticky notes, whiteboards, a mission to make AI clear |
| 2020 | Three founders, one small Bangalore room, endless chai |
| 2021 | A mighty team of five; first offsite plotted over a bonfire |
| 2022 | First real office; team of 18; first half-million in revenue |
| 2023 | I joined full-time as CTO — and eventually became co-founder; the great migration to a Bangalore HQ; crossed $1M in revenue |
| 2024 | Went global — planted a flag in the USA; 50+ strong |
| 2025 | Landed the first US client through the US business development office |
| 2026 | The year of the pivot — a deliberate reset that gave rise to Antesian Software Labs |
A company that grew the un-flashy way: a clear conviction, real clients, real solutions, revenue, and a team that compounded. Then, in 2026, the pivot. We ran, and we fell. In an environment that rewrites its own rules every few months, momentum isn't enough — you have to be willing to stop, take stock, and reset with intent. Out of that reset came a sharper bet for what comes next.
Pangea is where the lesson was learned: the model is rarely the bottleneck — context, trust, and the wiring around it are.
The conviction went with us. Antesian Software Labs is where it goes next — narrower by design, and in no rush to prove it.